Ethnic Conflict Regulation as Institutional Design:

The Case of the Western Balkans

Zoran Ilievski, University of Skopje

Stefan Wolff, University of Birmingham

ABSTRACT

This article develops a classification of institutions for the regulation of ethnic conflict by

focusing on three sets of challenges for institutional design in ethnically divided

societies: state construction, the institutions of governments, and the rights and

identities of groups and individuals. It examines existing prescriptions in three main

schools of ethnic conflict regulation (centripetalism, consociationalism, and power

dividing) in light of these institutional design challenges and contextualises them

empirically with reference to ethnic conflict regulation practice in the Western Balkans

over the past fifteen years. Finding a significant mix of institutions from across the three

theories of ethnic conflict regulation, the article proposes to conceptualise this practice

as ‘complex power sharing’ and proposes a research agenda to develop and test a more

comprehensive theory of complex power sharing.

2

I. INTRODUCTION

Relatively broad agreement exists within the literature on ethnic conflict regulation on

the utility of an approach that emphasises the design of (primarily political) institutions

as a means of accommodating competing claims of different conflict parties.1 Yet, this

literature at the same time is deeply divided: there is little agreement on what kinds of

institutions are best suited to regulating ethnic conflict. This division has its foundations

in different conceptions of the role and nature of ethnic identity, and especially its

malleability (or not), and consequently betrays a deep difference of opinion as to

whether institutions should recognise and accept what might appear as hardened and

politically salient ethnic identities or whether they should aim to transform the way in

which identities matter in politics and beyond. See from this perspective, the existing

literature on the subject offers both normative accounts of the desirability, and

empirical evidence of the feasibility, of designing institutional frameworks within which

disputes between different conflict parties can be accommodated, to such an extent

that political compromise becomes preferable to violent struggle, at least for the

majority of parties involved in the conflict in question. This debate about how to design

institutions to achieve sustainable peace in divided societies has engulfed the theory

and practice of ethnic conflict regulation for more than four decades, and has mainly

been fought between advocates of consociationalism and their opponents. The

disagreements between them have not subsided over the years and remain as divisive

1 Cf., for example, the contributions in Reynolds (2002), Choudhry (2008), Cordell and Wolff (2010), and Guelke and Tournon (2010). This agreement also extends to a rejection of partition as a means of regulating ethnic conflict: see O’Leary (2010), Sambanis (2000) and Sambanis and Schullhofer-Wohl (2009), but cf. Chapman and Roeder (2007) for evidence that partition fosters democracy in successor states. We exclude a discussion of partition here as it does not fall into the category of institutional approaches to ethnic conflict regulation.

3

as ever (cf., for example, Noel 2005; O’Flynn and Russell 2005; Taylor 2009). As a

consequence, the three main schools of thought on institutional design for ethnic

conflict regulation—centripetalism, consociationalism, and power dividing2—offer a

range of distinct prescriptions on how to ensure that differences of identity do not

translate into violence. In this sense, they are both realistic about the fact that

differences of identity exist and can create politically relevant and salient cleavages, and

optimistic about the possibility that such cleavages can be managed by accommodating

distinct identities and the demands they give rise to (cf. Horowitz 2002:19; McGarry and

O’Leary 2004b:8–12).

Disagreements between centripetalists, consociationalists, and power dividers,

while they have consequences for the prescription of institutional designs for ethnic

conflict regulation, do not, in fact, reflect the practice of conflict regulation very well. In

other words, power-sharing institutions remain the predominant feature of negotiated

agreements to end ethnic conflicts and are part of parcel of many constitutional reform

projects for the regulation of ethnic conflict before it erupts in to full-scale violence.3

This empirical reality is accepted by advocates of power dividing and centripetalism,

who do, however, add an immediate caveat as to whether power-sharing institutions

contribute to long-term peace and democracy. Thus, Rothchild and Roeder’s (2005)

main finding is that power sharing is a useful short-term mechanism to overcome

2 Using the labels centripetalism, consociationalism, and power dividing and concentrating on what we believe are the main proponents of particular approaches is but convenient shorthand for analytical purposes that glosses over the diversity of approaches within each school. 3 Hartzell and Hoddie (2008) make this point quite forcefully in their analysis of agreements negotiated to end civil wars. Wolff (2009) focuses more narrowly on ethnic conflicts, but includes cases short of civil war.

4

commitment problems4 that may prevent conflict parties in the immediate aftermath of

civil wars from agreeing to and sticking with a peace settlement, but that it is

detrimental in the long term, arguing that “*p+ower-sharing institutions *…+ in ethnically

divided societies after intense conflicts *…+ typically have a set of unintended but

perverse consequences” (Rothchild and Roeder 2005: 29). This is similar to Horowitz’s

(2007:1220) observation that “*c+ivil wars *…+ can sometimes be brought to an end with

consociational arrangements, but the desirability and durability of such agreements are

often in doubt.”

This raises two challenging theoretical as well as empirical questions for ethnic

conflict regulation: why do particular settlements (i.e., institutional designs) emerge,

and why do they succeed (or fail) in regulating ethnic conflict (in the sense of providing a

stable institutional framework in which conflict parties can address their differences

without recourse to violence)? Rather than attempting to answer these questions within

the limited space of a journal article, we take a step back and offer more detailed

reflection on the three distinct theories of conflict regulation in order to arrive at a

nuanced classification of institutional designs that is not confined within the existing,

and at times narrow and normatively pre-determined, theories of ethnic conflict

regulation. Starting with a discussion of the main challenges to institutional design for

ethnic conflict regulation, we examine centripetalism, power sharing, and power

dividing separately, summarizing the main tenets of each and paying particular attention

to their prescriptions for divided societies. Liberal consociationalism, centripetalism, and

4 On commitment problems, see Fearon (1995, 1998), Lake and Rothchild (1998), and Walter (2001).

5

power dividing thus serve as the theoretical framework for an empirical analysis of

ethnic conflict regulation qua institutional design in three countries in the Western

Balkans—Bosnia and Herzegovina, Macedonia, and Kosovo. Focusing on the distinct

dimensions of institutional design applied in each of the countries we rely

methodologically on a systematic comparison of relevant constitutional and other legal

acts that implement the relevant agreements in each of our cases.5 This enables us to

establish empirically the degree to which current ethnic conflict regulation theory and

practice overlap. The purpose of this exercise is not to demonstrate the inadequacy of

these theories individually or collectively, but to show that the reality of conflict

regulation points towards a high degree of complexity in providing a mixture of

recommendations from all three main theories. We conclude with some brief

observations on the failures and successes of conflict settlement in the Western Balkans

to date and suggest several avenues for further research into ethnic conflict regulation

theory.

II. ETHNIC CONFLICT REGULATION AND INSTITUTIONAL DESIGN

Advocating ethnic conflict regulation through institutional design assumes that such

conflicts can be resolved through an institutional bargain that establishes macro-level

5 The analysis is based on the following documents: “General Framework Agreement for Peace in Bosnia and Herzegovina” (www.intstudies.cam.ac.uk/centre/cps/documents_bosnia_dayton.html); “Brčko Final Award” (http://www.ohr.int/ohr-offices/brcko/arbitration/default.asp?content_id=5362; http://www.ohr.int/ohr-offices/brcko/arbitration/default.asp?content_id=5360); “Framework Agreement” http://www.intstudies.cam.ac.uk/centre/cps/documents_macedonia_frame.html) and “Law on Local Self-government of the Republic of Macedonia” (www.urban.org/PDF/mcd_locgov.pdf), The Comprehensive proposal for Kosovo Status Settlement, (http://www.unosek.org/unosek/en/statusproposal.html), Constitution of Kosovo, (http://www.kosovoconstitution.info/repository/docs/Constitution.of.the.Republic.of.Kosovo.pdf).

6

structures through which disputes among the conflict parties can be addressed

politically and without recourse to violence. This, however, is only element of a much

broader conflict regulation process, albeit a very central one. Focused essentially on the

outcome of negotiations, the institutional design approach is about finding the right

‘formula’ acceptable to the conflict parties (cf. Zartman 1989). The institutional bargain

achieved in negotiations is determined by a wide range of factors and in turn influences

the likelihood of the success of the conflict regulation process as a whole. This

relationship, however, is not linear, and questions of the quality of leadership of the

immediate conflict parties and of external diplomatic efforts to assist local leaders

throughout the negotiation, implementation, and operation phases of the conflict

regulation process co-determine eventual success or failure. With this caveat in mind,

we now turn to the challenges that institutional design faces in crafting macro-level

structures that regulate three contested areas in ethnic conflicts. An examination and

synthesis of the existing literature on ethnic conflict regulation6 suggests three different

dimensions of state construction:

(1) State construction, related particularly to questions of territorial structure; (2) The institutions of government, concerning among others the composition

and powers of the executive, legislative and judicial branches of government and the relationship between them; and

(3) Rights and identities of individual citizens and groups, i.e., the question if, and to what extent, individuals or groups are privileged.

6 Apart from the principal works and authors covered below, see also Bastian and Luckham (2003), Benedikter (2007), Choudhry (2008), Darby and McGinty (2003), Ghai (2000), Hechter (2000), Henrard (2000), Jarstadt and Sisk (2008), Lapidoth (1996), Noel (2005), Norris (2008), O’Flynn and Russell (2005), O’Leary, McGarry, and Salih (2005), Reynolds (2002), Schneckener and Wolff (2004), Taylor (2009), Weller and Metzger (2008), Weller and Wolff (2005), Wilford (2001), Woelk, Palermo and Marko (2008), and Wolff (2003). The synthesis of these and other sources is presented in the discussion in Sections II.1-3.

7

These three dimensions are inter-related and inter-dependent, but it is useful for

analytical purposes to keep them separate when exploring their specific aspects in some

degree of abstraction in the remainder of this section and in their practical

manifestations in Section IV.

II. 1. State Construction

The most important institutional design challenge in this area has to do with the

territorial organisation of the state. While the principal choice is generally between

unitary and federal systems, there is a great deal of variation within these two main

categories, and there are a number of hybrid forms as well. The most important

institutional design decision is about the number of levels of government with

substantive decision-making competences and the extent of these competences. Several

further decisions follow from this. The first one relates to the structural and functional

symmetry of the political-territorial organisation of the overall state. On the one end of

the spectrum, a state may be organised territorially in a completely symmetric fashion

with all territorial entities enjoying the exact same degree of functional competences,

exercising them through an identical set of local political institutions. However, the

nature of institutional design in divided societies may necessitate a different approach.

Thus, even where there is structural symmetry, functionally speaking the competences

enjoyed by different self-governing entities may differ, and/or they may exercise them

through different sets of political institutions. For example, where territorial sub-state

entities comprise ethnic groups distinct from that of the majority population, they may

8

be granted additional competences to address the particular needs of their

communities. In cases in which these sub-state entities are ethnically heterogeneous,

executive power sharing, reflecting local ethnic and political demographies, might be an

additional necessary feature of conflict resolution.

II. 2. The Composition and Powers of the Executive, Legislative and Judicial Branches of

Government and the Relationship between Them

The key aspects of institutional design in this area relate, first, to the nature of the

government system, i.e., whether it is a parliamentary, presidential or semi-presidential

system. A second dimension is the issue of whether executive and/or legislative power

sharing are mandatory, and if so, what the extent of prescribed inclusiveness is.

Inclusiveness, at the same time, is also an important feature of legislative design and is

primarily realised through the choice of an electoral system. Power sharing features and

inclusiveness may also extend into the judicial branch, primarily in relation to provisions

for the appointment of judges and prosecutors. A final issue in this regard is the overall

relationship between the three institutions of government, that is, the degree of

separation of powers between them. While this partially relates to the choice of

government system, it is also about the degree of independence of the judicial branch.

Institutional design thus not only prescribes certain outcomes in relation to the

composition of the executive, legislative and judicial branches of government but also

entrenches them in different ways from hard international law to domestic legislation.

9

II.3. The Relationship between Individual Citizens, Identity Groups and the State

Institutional design in this area is about the recognition and protection of different

identities by the state. On the one hand, this relates to human and minority rights

legislation, that is, the degree to which every citizen’s individual human rights are

protected, including civil and political rights, as well as the extent to which the rights of

different identity groups are recognised and protected. While there may be a certain

degree of tension between them, such as between a human rights prerogative of

equality and non-discrimination and a minority rights approach emphasising differential

treatment and affirmative action, the two are not contradictory but need to

complement each other in ways that reflect the diversity of divided societies and

contribute to its peaceful accommodation.

Moreover, the relationship between individuals, groups and the state is about

the degree to which institutional design favours particular groups and excludes others.

This is related to whether different groups are given different status (e.g., constituent

nations vs. minorities) and the political, economic and resource implications of this (e.g.,

mandatory inclusion in government, participation in proportional public sector job

allocation, reception of public funding, etc.). In other words, the question here is about

the degree to which specific group identities are recognised and protected and how this

manifests itself in the way in which the boundaries of authority are shaped by territory

or population groups.

10

III. INSTITUTIONAL DESIGN IN EXISTING THEORIES OF CONFLICT RESOLUTION7

Existing theories of conflict resolution generally acknowledge the importance and

usefulness of institutional design in conflict resolution, but offer rather different

prescriptions about what the most appropriate models are to achieve stable conflict

settlements. The three dominant theories in this respect are liberal consociational

power sharing, centripetalism, and power dividing. We discuss the main tenets of these

three schools of thought now in turn, focussing on their recommendations in each of

the three areas.

III.1. Liberal Consociationalism

Consociational power sharing is most closely associated with the work of Arend Lijphart,

who identified four structural features shared by consociational systems—a grand

coalition government (between parties from different segments of society), segmental

autonomy (in the cultural sector), proportionality (in the voting system and in public

sector employment) and minority veto (Lijphart 1977: 25-52). Consociationalism has

been developed further in the context of its use as a mechanism of interethnic

accommodation in Lijphart’s own later writings on the subject (e.g., Lijphart 1995,

2002), but more especially by John McGarry and Brendan O’Leary (McGarry 2006,

McGarry and O’Leary 2004a and b, O’Leary 2005a and b; see also Weller and Wolff

2005, and Wolff 2003, 2004). The most important modification of Lijphart’s original

theory is O’Leary’s contention that ‘grand coalition’ (in the sense of an executive

7 The following section draws on Co-author (2010).

11

encompassing all leaders of all significant parties of all significant communities) is not a

necessary criterion. Rather, O’Leary demonstrates that what matters for a democratic

consociation ‘is meaningful cross-community executive power sharing in which each

significant segment is represented in the government with at least plurality levels of

support within its segment’ (O’Leary 2005a: 13).8

The scholarly literature on consociationalism distinguishes between corporate

and liberal consociational power sharing, the latter now the more common policy

prescription among consociationalists.9 The main difference between the two is that a

‘corporate consociation accommodates groups according to ascriptive criteria, and rests

on the assumption that group identities are fixed, and that groups are both internally

homogeneous and externally bounded’, while ‘liberal … consociation … rewards

whatever salient political identities emerge in democratic elections, whether these are

based on ethnic groups, or on sub-group or trans-group identities’ (McGarry 2006: 3,

see also Lijphart 1995 and O’Leary 2005a).

Territorial self-governance is a significant feature within the liberal

consociational approach which, in this context, emphasises that the self-governing

territory should define itself from the bottom up, rather than be prescribed top-down.10

8 On this basis, O’Leary (2005a: 12-13) distinguishes between three sub-types of democratic (i.e., competitively elected) consociation: complete (executive composed of all leaders of all significant segments), concurrent (all significant segments represented, and executive has at least majority support in all of them), and weak (all significant segments represented, and executive has at least one segmental leadership with only plurality support). 9 Corporate consociationalism, however, is still evident to some extent in political practice: for example, Bosnia and Herzegovina, under the original Dayton Accords, Northern Ireland under the 1998 Agreement, Lebanon under the National Pact and under the 1989 Ta’if Accord, Cyprus under the 1960 constitution and proposed (but rejected) Annan Plan display features of pre-determined arrangements based on ascriptive identities. 10 In the context of Iraq, McGarry (2006: 6-7) explains how this process has been enshrined in the Iraqi constitution: “Kirkuk can choose to join Kurdistan if its people want. Governorates in other parts of the country are permitted to amalgamate, forming regions, if there is democratic support in each governorate. In this case, a twin democratic

12

Liberal consociationalists consider arrangements in which there are more than two, and

ideally even more than three, self-governing entities within a given state, as conducive

to the chances of state survival. Liberal consociationalists equally support the principle

of asymmetric devolution of powers, i.e., the possibility for some self-governing entities

to enjoy more (or fewer) competences than others, depending on the preferences of

their populations (cf. McGarry 2007).

Naturally, self-governance is complemented with what liberal consociationalists

term ‘shared rule’, i.e., the exercise of power at and by the centre across the state as a

whole. While the other three key features of Lijphartian consociationalism (apart from

‘segmental autonomy’) continue to be favoured by liberal consociationalists, such as

grand coalitions, proportionality and minority veto rights, the emphasis is on

cooperation and consensus among democratically legitimised elites, regardless whether

they emerge on the basis of group identities, ideology or other common interest. Liberal

consociationalists thus favour parliamentary systems,11 proportional (PR list) or

proportional preferential (STV) electoral systems, decision-making procedures that

require qualified and/or concurrent majorities, and have also advocated, at times, the

application of the d’Hondt rule for the formation of executives12 (cf. Lijphart 2004,

O’Leary 2005a, see also Wolff 2003).

threshold is proposed: a vote within a governorate’s assembly and a referendum. … It is also possible for Shi’a dominated governorates that do not accept SCIRI’s vision to remain separate, and, indeed for any governorate that may be, or may become, dominated by secularists to avoid inclusion in a sharia-ruled Shiastan or Sunnistan.” 11 Note, however, that, empirically, collective presidential systems are as widespread in existing functioning consociations than parliamentary ones. Personal communication from Brendan O’Leary. 12 For details on the d’Hondt rule, see O’Leary, Grofman and Elklit (2005).

13

This means, liberal consociationalists prefer what O’Leary refers to as ‘pluralist

federations’ in which co-sovereign sub-state and central governments have clearly

defined exclusive competences (albeit with the possibility of some concurrent

competences) whose assignment to either level of authority is constitutionally and,

ideally, internationally, protected, in which decision-making at the centre is consensual

(between self-governing entities and the centre, and among elites representing different

interest groups), and which recognise, and protect the presence of different identities

(O’Leary 2005b).

In order to protect individuals against the abuse of powers by majorities at the

state level or the level of self-governing entities, liberal consociationalism offers two

remedies—the replication of its core institutional prescriptions within the self-governing

entity,13 and the establishment and enforcement of strong human and minority rights

regimes both at the state and sub-state levels. In addition, the rights of communities—

minorities and majorities alike—are best protected in a liberal consociational system if

its key provisions are enshrined in the constitution and if the interpretation and

upholding of the constitution is left to an independent and representative constitutional

court whose decisions are binding on executive and legislature (cf. O’Leary 2005b: 55-8).

III.2. Centripetalism

Centripetalism emphasizes that rather than designing rigid institutions in which elected

representatives have to work together after elections, “intergroup political

13 On regional consociations see Wolff (2004).

14

accommodation” is achieved by “electoral systems that provide incentives for parties to

form coalitions across group lines or in other ways moderate their ethnocentric political

behaviour” (Horowitz 2004:507–8). This school of thought is most prominently

associated with the work of Donald Horowitz (1985; 1990; 1991; 2002), as well as with

that of Timothy D. Sisk (1996), who uses the terms “integrative” and “integration” when

referring to centripetalism (as do Rothchild and Roeder 2005b: 35), Matthijs Bogaards

(1998; 2000; 2003), who initially criticized consociationalism on conceptual and

methodological grounds (Bogaards 1998; 2000), before offering a strongly centripetal

alternative (Bogaards 2003), Benjamin Reilly (1997; 2001; 2006), and Andreas Wimmer

(2003). Reilly, for example, advocates, among others, ‘(i) electoral incentives for

campaigning politicians to reach out to and attract votes from a range of ethnic groups

other than their own…; (ii) arenas of bargaining, under which political actors from

different groups have an incentive to come together to negotiate and bargain in the

search for cross-partisan and cross-ethnic vote-pooling deals…; and (iii) centrist,

aggregative political parties or coalitions which seek multi-ethnic support…’ (Reilly 2001,

p. 11; emphasis in original). This is partially echoed by Wimmer in his proposals for the

first post-Saddam Iraqi constitution to introduce ‘an electoral system that fosters

moderation and accommodation across the ethnic divides’, including a requirement for

the ‘most powerful elected official … to be the choice not only of a majority of the

population, but of states or provinces of the country, too’, the use of the alternative

vote procedure, and a political party law demanding that ‘all parties contesting elections

… be organised in a minimum number of provinces’ (Wimmer 2003). In addition,

15

Wimmer advocates non-ethnic federalism (ibid.: 123-5), at least in the sense that there

should be more federal entities than ethnic groups, even if a majority of those entities

would be more or less ethnically homogeneous or be dominated by one ethnic group.

Furthermore, “a strong minority rights regime at the central level, a powerful

independent judiciary system and effective enforcement mechanisms are needed”,

according to Wimmer (2003: 125).

In what remains a classic work in the field of ethnic conflict and conflict

resolution theories, Donald L. Horowitz (1985 [2000]) discusses a range of structural

techniques and preferential policies to reduce ethnic conflict. Among them, he

emphasises that ‘the most potent way to assure that federalism or autonomy will not

become just a step to secession is to reinforce those specific interests that groups have

in the undivided state’ (Horowitz 1985 *2000+: 628). Horowitz also makes an explicit

case for territorial self-governance (i.e., federalism) in his proposals for constitutional

design in post-apartheid South Africa (Horowitz 1991: 214-226) and argues, not

dissimilar to power dividing advocates, for federalism based on ethnically

heterogeneous entities. In a later study, more explicitly focused on federation as a

mechanism for conflict reduction, Horowitz (2007) accepts that homogeneous

provinces, too, can prove useful for this purpose, but argues that rather than the aim

being to facilitate group autonomy (the consociational rationale), homogeneous

provinces offer the possibility to foster intra-group competition (2007:960–1; see also

Horowitz 2008:1218). In an earlier contribution to the debate, Horowitz had recognized

the need for federal or autonomy provisions, but cautioned that they could only

16

contribute to mitigating secessionist demands if “*c+ombined with policies that give

regionally concentrated groups a strong stake in the center” (1993:36). Interestingly,

however, this need for centripetal elements in territorial designs for conflict resolution

is also echoed in some corners of the consociational school (cf. Weller and Wolff 2005).

Similar to Wimmer (2003; see above), Horowitz, citing the Nigerian experience, sees

utility in splitting large ethnic groups into several provinces as this potentially

encourages the proliferation of political parties within one ethnic group, resulting in

intra-group competition and a lessened impact of relative numerical superiority of one

group over others (Horowitz 2000:602–4; 2007:960–1; see also 2008:1218).

While centripetalism is thus open to engaging with, among others, territorial

approaches to conflict settlement, “its principal tool is *…+ the provision of incentives,

usually electoral incentives, that accord an advantage to ethnically based parties that are

willing to appeal, at the margin and usually through coalition partners of other ethnic

groups, to voters other than their own (Horowitz 2008:1217, our emphasis). In

particular, Horowitz emphasizes the utility of electoral systems that are most likely to

produce a Condorcet winner, i.e. a candidate who would have been victorious in a two-

way contest with every other candidate in a given constituency. The most prominent

such electoral system is the alternative vote (AV), a preferential majoritarian electoral

system, that is said to induce moderation among parties and their candidates as they

require electoral support from beyond their own ethnic group in heterogeneous, single-

seat constituencies (Horowitz 2003:122–5).

17

III. 3. Power Dividing

In the context of conflict resolution, the theory of power dividing has been put forward

most comprehensively by Philip G. Roeder and the late Donald Rothchild in their co-

edited volume Sustainable Peace: Power and Democracy after Civil Wars (Roeder and

Rothchild 2005). Power dividing is seen as “an overlooked alternative to majoritarian

democracy and power sharing” as institutional options in ethnically divided societies

(Rothchild and Roeder 2005: 6). Three strategies that are said to be central to power

dividing—civil liberties, multiple majorities, and checks and balances—in practice result

in an allocation of power between government and civil society such that “strong,

enforceable civil liberties … take many responsibilities out of the hands of government”,

while those that are left there are distributed “among separate, independent organs

that represent alternative, cross-cutting majorities”, thus “balanc*ing+ one

decisionmaking centre against another so as to check each majority … *f+or the most

important issues that divide ethnic groups, but must be decided by a government

common to all ethnic groups” (Rothchild and Roeder 2005: 15).

The key institutional instruments by which power dividing is meant to be realised

are, first of all, extensive human rights bills that are meant to leave “key decisions to the

private sphere and civil society” (Rothchild and Roeder 2005: 15). Second, separation of

powers between the branches of government and a range of specialised agencies

dealing with specific, and clearly delimited, policy areas are to create multiple and

changing majorities, thus “increas*ing+ the likelihood that members of ethnic minorities

will be parts of political majorities on some issues and members of any ethnic majority

18

will be members of political minorities on some issues” (Rothchild and Roeder 2005: 17).

Third, checks and balances are needed “to keep each of these decisionmaking centres

that represents a specific majority from overreaching its authority” (ibid.). Thus, the

power dividing approach favours presidential over parliamentary systems, bicameral

over unicameral legislatures, and independent judiciaries with powers of judicial review

extending to acts of both legislative and executive branches. As a general rule, power

dividing as a strategy to keep the peace in ethnically divided societies requires

“decisions *that+ can threaten the stability of the constitutional order, such as

amendments to peace settlements” be made by “concurrent approval by multiple

organs empowering different majorities” (Rothchild and Roeder 2005: 17).

III. 4. The Different Theories Compared

The preceding overview of three main theories of conflict resolution illustrates two

important aspects of current academic and policy debates about how to establish

sustainable institutional settlements in cases of self-determination conflicts: while there

are fundamental differences in the underlying assumptions about how such settlements

can succeed, certain institutional arrangements that complement the basic prescriptions

of each approach are largely similar, if not identical (see Table 1).

<<Table 1 about here>>

19

IV. INSTITUTIONAL DESIGN IN PRACTICE: AN EMPIRICAL ANALYSIS OF BOSNIA AND HERZEGOVINA,

MACEDONIA, AND KOSOVO

The selection of Bosnia and Herzegovina, Macedonia and Kosovo as our cases for

empirical illustration of ethnic conflict regulation qua institutional design has several

reasons. Individually the cases have received considerable attention, even though the

amount of coverage has differed from case to case with Bosnia and Herzegovina the

most widely covered, followed by Macedonia, and Kosovo. The analysis of Bosnia and

Herzegovina is quite comprehensive in terms of all aspects of the conflict regulation

process, this is less the case for Macedonia and certainly for Kosovo where more

attention has been focused on the process of negotiating than on the specific

institutional settlement adopted as a result. Our analysis will thus fill a gap in the

existing literature by offering a systematic comparison of the institutional design

features of the settlements currently in place. Moreover, while an examination of the

origins or consequences of the three institutional designs in place is not the primary

purpose of our analysis, the comparison is more relevant as the broader context in

which institutions were adopted is broadly similar in all cases in several crucial

dimensions, including the fact that settlements were arrived at after violent conflict

(albeit at different scales) and with significant international involvement before, during,

and after settlement conclusion (albeit by different actors). On the other hand, as we

will demonstrate below, the nature of each settlement is quite different, thus

collectively offering a wider range of available options of institutional design. The three

cases thus serve as a test of our broader framework of classifying institutional designs

20

for ethnic conflict regulation along the three dimensions of state construction,

institutions of government, and rights and identities.

The dimensions of the comparative analysis flow directly from the discussion of

institutional design and the examination of the three approaches to conflict resolution

above. Specifically, the comparative analysis of state construction will focus on two

aspects:

symmetry and asymmetry in institutional design; and

the distribution of powers between different levels of government.

As far as the institutions of government are concerned, we will consider five separate

issues:

the nature of the government system;

the choice of the electoral system;

power sharing in executive and/or legislature;

the degree of independence of the judiciary; and

legal entrenchment.

Finally, when examining rights and identities, two dimensions are of particular

importance:

human and minority rights provisions; and

recognition and protection of identities.

21

IV.1. State Construction

Symmetry and Asymmetry in Institutional Design14

The first element to consider in the context of questions about symmetry and symmetry

of institutional design is the number of levels of government that actually exist across

our three case studies (Table 2).

<<Table 2 about here>>

In the case of Bosnia and Herzegovina more than three levels of government exist. This

is a result of the interplay of domestic (i.e., state and sub-state), regional and

international factors in the process of state creation at Dayton, leading to a complex

federal-confederal structure of the state. Yet, Dayton proved inconclusive in relation to

one territorial aspect that required subsequent arbitration between Republika Srpska

and the the Federation of Bosnia and Herzegovina: the District of Brčko has a special

status as a territory held in condominium and belonging simultaneously to the two

Entities (Republika Srpska and Federation of Bosnia and Herzegovina) but self-governing

as a unitary territory in which neither Entity exercises any authority.

In Macedonia and Kosovo, structural state construction is far less complicated

than in the case of Bosnia and Herzegovina and only two levels of government exist—

central government and local governments. Both are prescribed in the relevant

constitutions, and their functions and powers are detailed there and in relevant

legislation. In Macedonia, a legally guaranteed opportunity for citizens exists to develop

a further level of government at the level of neighbourhoods, but this is regulated by by-

14 For a discussion of the utility of asymmetric designs for conflict resolution, see McGarry (2007).

22

laws of the individual local governments and thus a matter of local decision-making

rather than of state construction. In Kosovo, too, an additional provision is of interest,

namely the opportunity of local government entities to cooperate on matters of

common concern. This is (potentially) particularly relevant for Serb-majority

communities in the north, and we discuss it in greater detail in section III.3.

Another way of looking at different levels of government is to examine the

degree to which they represent structurally and/or functionally symmetric or

asymmetric institutional designs (Table 3),15 as this perspective provides a more

comprehensive picture of state construction.

<<Table 3 about here>>

Given the small number of cases involved, there is little point here in attempting to

detect a general pattern of symmetry or asymmetry. What is noteworthy, however, is

that while symmetric structures and symmetric functions may be correlated

(Macedonia), symmetric structures do not preclude asymmetric functional capacities

(Kosovo). From a theoretical point of view, it is worth pointing out that both liberal

consociational power sharing and centripetalism, albeit to differing degrees, allow for

asymmetric structures and functions. While liberal consociational power sharing is

principally in favour of territorial configurations reflecting the expressed wishes of self-

defined communities (whatever the basis of such self-definition), centripetalism is not

opposed to the use of territorial self-governance arrangements in either symmetric

15 Structural asymmetry is meant to signify the existence of territorial entities that do not ‘fit’ the overall construction of the state, e.g., an autonomous territory in an otherwise unitary state, or, as in the case of Bosnia and Herzegovina, the existence of a special-status district within a confederation. Functional asymmetry is meant to signify that some territorial entities enjoy a different measure of competences, e.g., have wider legislative powers than others.

23

(federalism) or asymmetric (autonomy) forms, but crucially in this respect,

centripetalists and advocates of power dividing prefer territorial self-governance to be

based on ‘administrative’ rather than ‘ethnic’ criteria, in an effort to prevent the

institutionalisation of group identities and enable coalitions of interest based on policy

rather than identity (centripetalists) or multiple and changing majorities (power

dividers). Having said that, it is evident that in the cases that form the basis of this

empirical comparison the entities of territorial self-governance are exclusively those in

which group identities form the basis of boundaries. Having said that, the territorial

organisation of Kosovo and Macedonia with its unitary state structure and devolution of

powers to local governments rather than to an intermediary ‘regional’ level is more in

line with one recommendation of centripetalists, namely to split up compact ethnic

communities into several territorial entities rather than creating a single self-governing

entity for them. While liberal consociationalists do not specifically recommend this, they

are nonetheless aware of the poor track record of two- or three-entity federations and

caution against such designs where they can be avoided. Regarding Kosovo, it is also

noteworthy that the allocation of enhanced competences to municipalities with local

Serb majorities is a novel feature in institutional design: while avoiding formal regional

autonomy for these areas, it nonetheless sets them aside from ‘ordinary’ municipalities

and accommodates ethnic differences flexibly at the local level.

24

Distribution of Powers

One of the key questions to ask of governance regime is where powers rest; i.e., how

different competences are allocated to different levels of government and whether they

are their exclusive domain or have to be shared between different levels of government.

As with other dimensions in this analysis, there is a certain degree of context-dependent

variation across the cases under examination. To begin with the allocation of powers to

different levels of government, in Bosnia and Herzegovina the state level has retained

very few powers which are specifically mentioned in the Dayton constitution, including

foreign and foreign trade policy, customs policy, monetary policy, immigration, criminal

law enforcement, international communications facilities, inter-Entity transportation,

and air traffic control. All other powers, by default, fall to the Entities and the District of

Brčko, respectively. In other words, residual authority is, unusually, not held by the

centre but by sub-state entities. This is further complicated by the fact that power is

further devolved within the Federation to the cantons in a similar manner, limiting the

powers of the Federation to a few specifically mentioned in the Federation constitution

(including military and defence matters, citizenship, economic and fiscal policy, inter-

cantonal crime, and energy policy), and assigning all further competences either to a

joint list (including human rights, health, environment, infrastructure, social welfare,

tourism, and use of natural resources) or by default exclusively to the cantons. This

replicates the State-level principle of assigning residual authority to the lower level of

government.

25

The situation in Bosnia and Herzegovina stands in clear contrast to both Kosovo

and Macedonia. Here, the centre holds residual authority and competences are

specifically assigned to local governments, including in the areas of urban planning,

communal activities, culture, sport, social security and child care, pre-school and

primary education, and basic health care.

In Kosovo, powers devolved to the local level are wide-ranging and include local

economic development; urban and rural planning, provision and maintenance of public

services and utilities, public pre-primary, primary and secondary education, promotion

and protection of human rights, public health and health care, social welfare services,

public housing, naming of roads, streets and other public places. Specifically, the law on

local government also gives municipalities power on any matter which is not explicitly

excluded from their competence nor assigned to any other authority. From this

perspective, municipalities in Kosovo hold residual authority.

Despite its character as a unitary state, some municipal governments in Kosovo,

principally those with an ethnic Serb majority, can exercise so-called enhanced

municipal competences in the areas of health, education and cultural affairs. Moreover,

they are given participatory rights in selecting local station police commanders.16 The

state, however, retains the right to monitor the exercise of these powers along clearly

specified criteria, including equal access to public services, quality and quantity

standards, and minimum standards for accreditation of public service providers.

16 Whereas in the rest of Kosovo, the General Director of the Kosovo Police Service presents three candidates for the post to the municipal assembly for ranking and subsequent appointment by the General Director, in communities where the largest ethnic community is Serb, the municipal assembly initiates the process by proposing two candidates for selection and appointment by the Ministry of Internal Affairs.

26

Moreover, as noted earlier, these municipalities may also cooperate with any other

municipality in providing services. Specifically, the municipalities of Mitrovicë/Mitrovica

North, Graçanicë/Gracanica, Shtërpcë/Štrpce also have powers in secondary health care

and Mitrovicë/Mitrovica North has enhanced competences in relation to university

education, including registration and licensing, recruitment, payment of salaries and

training of education instructors and administrators. Kosovo Serb majority municipalities

anywhere in Kosovo furthermore enjoy a range of enhanced competences in the area of

culture, including religious and cultural heritage.

As already indicated, variation also exists primarily with regard to the way in

which powers are allocated and the degree of flexibility concerning new fields of policy-

making not relevant or not included at the time a specific agreement was concluded.

The principal mechanism to handle the distribution of powers is the drawing up of lists

that enumerate precisely which powers are allocated to which levels of government

and/or which are to be shared between different such levels. These lists can be very

specific for each level of government, they can be specific for one or more levels and

‘open-ended’ for others, or they can be mixed in the sense that certain powers are

exercised jointly or concurrently by different levels of government. The key issue in any

of these arrangements is which level of government retains residual authority for any

partly devolved power or any other policy area not explicitly allocated elsewhere (see

Table 4).

<<Table 4 about here>>

27

None of the three theories of conflict resolution discussed above offers much specific

guidance on this issue of power allocation to different vertical levels of government.

Some inferences can nevertheless be made. Power-dividers, who express a certain

preference for the American model of federalism, favour strong central governments

and are thus likely to opt for residual authority to remain with the central government.

A similar tendency can be observed for advocates of centripetalism (cf., for example,

Wimmer 2003). For the liberal consociational school of power sharing, it is important

that power sharing is a more attractive option to conflict parties than recourse to

violence, hence its advocates should be interested in substantive powers assigned to

territorial self-government entities, which is best done either by way of assigning

residual authority to these entities or by drawing up specific lists which clearly distribute

powers between central and sub-state governments.

IV.2. The Composition and Powers of the Executive, Legislative and Judicial Branches of

Government and the Relationship between Them

The Nature of the Government System and the Choice of Electoral Systems

All three approaches to conflict resolution strongly advocate the separation of powers

between executive, legislature and judiciary, and this has been achieved in the

institutional designs of all cases discussed here. In particular, across all three country

cases independent judicial systems exist, including also at the level of the Entities and

the District of Brčko in Bosnia and Herzegovina. Beyond this general point, significant

variation remains (see Tables 5.1-2).

28

At the State level, Bosnia and Herzegovina is a semi-presidential system with a

collective presidency with rotating chairmanship. The Presidency consists of three

members: one Bosniak and one Croat, each directly elected from the territory of the

Federation, and one Serb directly elected from the territory of the Republika Srpska. The

electoral system for this is a simple majority vote with a second-round run-off between

the two candidates topping the first-round ballot if no candidate achieves 50%+1 of the

vote in the first round. At the sub-state level, the Entities and the District of Brčko are

parliamentary systems, electing their executives in parliament. The Republika Srpska

also has a directly elected, collective presidency with a president and two vice

presidents, each elected from a different constituent people, but with only symbolic

powers. While the State and the Federation have bicameral parliaments, legislatures in

the Republika Srpska and the District of Brčko are unicameral.

In the same way in which the variation of government systems in Bosnia and

Herzegovina derives from the complex multilevel structure of its institutions, the

comparative simplicity of the respective systems in Macedonia and Kosovo—unicameral

and parliamentary—is by virtue of the unitary structure of their states. The only

difference between the two is that the Macedonian president is elected directly (using

the same two-round run-off system as in Republika Srpska) while the president of

Kosovo is elected by parliament. Both have all but symbolic powers.

In all three cases, the central legislature is elected by List PR systems. In Bosnia

and Herzegovina, the lower chamber of parliament has 42 members which are elected

by a PR-List system, albeit one that uses separate electoral roles for the Federation of

29

Bosnia and Herzegovina (28 seats) and the Republika Srpska (14 seats). Kosovo also uses

a PR-List system, but one that is qualified in two important dimensions. Since the

November 2008 elections, an open list system is used allowing voters to choose up to

ten candidates from different lists. Moreover, for the first two electoral terms under the

new constitution 20 seats in the 120-member parliament are reserved for Serbs (10) and

other ethnic groups (10). The 120 members of the Macedonian parliament are elected

through PR-List system in 6 electoral districts, each electing 20 members.

<<Tables 5.1-2 about here>>

Thus, as far as the nature of the government system and the choice of electoral systems

are concerned, we see a clear dominance of (liberal) consociational views: in all but one

case, the system of government is parliamentary, and in the one case where it is semi-

presidential instead, the presidency is a collective one with a rotating chairmanship.

Moreover, the predominant electoral system is List-PR, a system preferred by Lijphart

and shown by Norris (2008) to lead to sustainable power sharing arrangements.

Power Sharing

One element of the complexity of the institutional designs in the three country cases

discussed here stems from the fact that constitutional engineers have developed

innovative ways to combine traditional structures of horizontal power sharing and

vertical power dividing; in other words, power is often shared at different levels of

government, according to different rules, and using various arrangements. Table 6.1

presents an overview of where power is shared in our cases, while Table 6.2 gives an

30

indication of how it is shared—qua mandatory representation in executive and/or

legislature or through ensuring cross-community consensus by requiring qualified

majorities for the adoption of decisions in legislatures.

<<Tables 6.1-2 about here>>

As can be expected, Bosnia and Herzegovina has the most extensive power sharing

arrangements in terms of where power is shared (State, Entities, Cantons, and District of

Brčko) and how (mandatory representation of relevant constituent peoples in

legislatures, executives, and presidencies, qualified majority voting on specific ‘vital

interest’ issues in State, Federation, and District of Brčko legislatures).17 This reflects not

merely ethnic diversity across different levels of government throughout the country

but also the recognition that mere devolution of powers to a lower level of government

without ensuring proper safeguards against majoritarian abuse of such power would

simply replicate the conflict between different groups at the state level where power

sharing has provided some sort of a solution, albeit an imperfect one.

In principle, this state of affairs is replicated in Kosovo. The 2008 constitution

foresees the mandatory representation of non-Albanian communities in the state-level

executive and legislature and in municipal assemblies and executives and requires

parallel consent in the Kosovo assembly (i.e., from a majority overall and from a

majority of representatives from non-majority communities) on vital-interest legislation.

According to the Constitution, power sharing at the central level, as well as in

municipalities where at least 10 % of the population is non-Albanian is required.

17 Moreover, specific arrangements also exist in many other ethnically heterogeneous localities across the Federation.

31

According to Art. 96 of the Constitution of Kosovo, the Serb community, as well as other

Kosovo non-majority communities are entitled to a minister and two deputy ministers in

the Government of Kosovo. Guaranteed seats are available for the Serbs and other non-

Albanian communities in the Assembly, the Judicial Council and in the procedure of

electing judges in the minority populated areas.18 At the municipality level, there are

positions of vice presidents foreseen in each of municipality where at least 10 % of the

population is non-Albanian.19 However, in practice the entire power sharing system in

Kosovo is of limited effectiveness at the moment as Serbs from the northern Mitrovica

region boycott the Kosovo political institutions at the moment and have, with the

support and encouragement of Belgrade, established comprehensive parallel

structures.20

In contrast, there are no provisions for mandatory representation in executive or

legislative bodies in Macedonia either at the centre or at lower levels of government.

Yet, the absence of formal structures of power sharing at the centre does not preclude

power nevertheless being shared to some extent as the country’s demographic

balances, structure of the party system and electoral formula combine in a way that

have made the formation of government coalitions between ethnic Macedonian and

ethnic Albanian parties the norm, and ethnic Albanian parties have been present in all

coalition governments since Macedonia’s independence. Even the 1990-1992

“government of experts”, which was not structured around political parties, included

18 Articles 64, 103 and 108 of the Constitution of Kosovo. 19 Article 62 of the Constitution of Kosovo. 20 For a comprehensive overview see: OSCE Report, Parallel Structures in Kosovo 2006-2007, <http://www.osce.org/documents/mik/2007/04/23925_en.pdf >

32

three ethnic Albanians. Moreover, the 2001 Ohrid Agreement specified several vital-

interest areas in which parallel consent applies for the legislature to pass any laws thus

further ensuring meaningful participation of Albanians at the centre. The “Badinter

principle”, i.e., parallel consent, is also enshrined in the Law on Local Self-government

(2002), establishing a form of legislative power sharing in the areas of culture, language

and symbols at the municipality level Municipalities with over 20% minority population

also have Committees on Inter-ethnic Relations. In addition, the numerical strength and

concentration of ethnic Albanians in western Macedonia allows them to benefit fully

from the implementation of local autonomy as foreseen in the Ohrid Agreement, while

the geographical concentration and size of the minority make a solution based on

substantive regional autonomy less attractive for ethnic Macedonians, as it could be

construed as a first step to the partition of the country.

In all three countries, and across different levels of government, we find power

sharing arrangements in place, manifesting themselves in the representation of

members of different communities in executive and legislative bodies and in their

meaningful participation in decision making qua qualified majority voting procedures

(super majorities or parallel consent). While the voluntary emergence of power sharing

governments at the central level in Macedonia could be interpreted as in line with the

centripetal approach, neither the electoral system used nor the application of qualified

majority voting procedures in parliament and in the local self-government is in a strict

sense. However, the principle of qualified majority voting does not give a strict veto

right to every community per se, but rather enables individual members of different

33

minorities sitting in legislatures, as well as different parties representing the same

minority in a legislative body to form different coalitions depending on the decisions in

question. In this sense, we also find a mechanism in this principle that creates a

centripetalist ‘arena for bargaining’ and enables the kind of issue-based changing

majorities that power dividers recommend (albeit within one rather than many decision-

making bodies). Equally importantly, power sharing arrangements in Kosovo, and even

more so in Bosnia and Herzegovina, display many features of corporate rather than

liberal consociationalism. Principles of power dividing, where they have been applied,

are only relevant to the extent that they are more generally embraced as principles of

state construction, such as in terms of the separation of powers.

Legal Entrenchment

Guarantees of institutional structures of horizontal and vertical power sharing and

power dividing are essential to prevent the arbitrary abrogation of devolved powers and

thus to ensure conflict parties of the relative permanence of the institutions they agreed

upon. Guarantees are particularly important for the relatively weaker party in a self-

determination dispute, i.e., a specific minority, to protect it from a state reneging on

earlier concessions. However, such guarantees are also valuable for states in that they

commit all parties to an agreed structure and, in most cases, imply that there can be no

unilateral change of recognised international boundaries outside pre-determined

procedures.

34

In principle, guarantees can be either international or domestic, and in the latter

case they can be part of a country’s constitution or other legislation (see Table 7). Given

the complexity of many of today’s self-determination conflicts, guarantees often exist at

more than one level. In addition, international guarantees can take the form of hard

guarantees (international treaties) or of ‘soft’ guarantees (non-binding standards and

norms, declarations of intent, etc.).

<<Table 7 about here>>

Table 7 illustrates that there is variance across the cases considered here. In terms of

the strength of the protection that they afford established horizontal and vertical power

sharing and power dividing structures, hard international guarantees are preferable

over other forms of guarantees, provided there is significant commitment of the

international community to uphold its guarantees. In Bosnia and Herzegovina this

commitment is unquestionable with the presence of peacekeeping forces in both

territories and with the investment that has been made over the past years by the

international community in order to foster economic development, and institution-

building and reform.

In Bosnia and Herzegovina, international bodies with a clear mandate exist: the

multi-national Peace Implementation Council and the UN Security Council, respectively.

As discussed above, the Office of the High Representative is invested with significant

powers to intervene into the political process in Bosnia and Herzegovina at all levels.

Specifically for the District of Brčko, an International Supervisor with similar powers

exists.

35

Regarding international guarantees in Kosovo, the Ahtisaari plan was accepted

by the Assembly of Kosovo in spring 2008 (in the absence of a Security Council

resolution). The Constitutional Provisions required by the Ahtisaari package were

included in the Kosovo Constitution. In addition, besides the laws that were brought

before the proclamation of independence in February 2008, in which special rights for

Kosovo communities are guaranteed, after the proclamation of independence almost

forty so called “Ahtisaari-required laws” (Annex 12 of the Ahtisaari package) were

approved by the Kosovo Assembly. As in Bosnia and Herzegovina, a heavy international

presence remains in Kosovo. The UN still plays a role based on Security Council

resolution 1244(1999), although its role is continuously decreasing in favour of the

International Civilian Office (ICO), which is now responsible for monitoring the

implementation of the Ahtisaari plan. The International Civilian Representative (ICR) in

Kosovo is the final authority for the interpretation of the Ahtisaari Plan and its

implementation. As well as approving all judges and prosecutors selected, and

appointing and approving key positions in Kosovo’s institutions, the ICR has executive

corrective powers. These powers include the annulment of laws or decisions taken by

the Kosovo Assembly if he decides they breach or are inconsistent with the Ahtisaari

Settlement, or undermine the rule of law. The ICR can also sanction or remove from

office any public official if these officials have opposed the letter or spirit of the

Settlement, and/or if they have obstructed the work of the ICR or EULEX.

In contrast to these hard international guarantees in place in Kosovo and Bosnia

and Herzegovina, soft international guarantees primarily manifest themselves in the

36

form of the involvement of international organisations in the negotiation,

implementation, and (potentially) operation of a particular peace agreement. While not

of the same legally binding and thus potentially enforceable status as hard international

guarantees, a significant presence of international agents is often instrumental in

shaping preference and opportunity structures for the conflict parties. In the cases

studies, this has taken different forms. In all three cases, an international military and

civilian presence, as well as the involvement of various international governmental and

non-governmental organisations on an unprecedented scale, have, for better or worse,

been instrumental in the implementation and operation of the respective agreements

thus far.

At the level of domestic guarantees, constitutional guarantees are more

entrenched than those which have their source in normal legislation. Incorporation of

specific provisions of peace agreements into constitutions is a common way of realising

constitutional guarantees and has occurred in Bosnia and Herzegovina, Macedonia, and

Kosovo.

All three theories of conflict resolution emphasise the importance of judicial

institutions and enforcement mechanisms to uphold the letter and spirit of conflict

settlements. They also recognise the role of external actors and guarantees, but are

divided on the usefulness of international intervention: power-dividers see a limited,

transitional role for them; advocates of both liberal consociationalism and

centripetalism embrace them more readily as facilitators and guarantors of settlements.

However, it is worth noting that a key difference remains between the two approaches.

37

Centripetalists consider that their institutions will be self-sustaining after they are set up

with outsiders’ help, and that consociational institutions will need to be constantly

propped up by external third parties. Liberal consociationalists, by contrast, take their

preferred arrangements to be much more likely to be acceptable to local elites and

therefore less dependent on external imposition than centripetalist or power dividing

institutional designs.

IV.4. The Relationship between Individual Citizens, Identity Groups and the State

Human and Minority Rights Provisions

Relevant human and minority rights provisions exist in all cases included in this analysis,

albeit to differing degrees. Applicable law includes international and regional standards

and more specific state-wide, and in some cases local, human and minority rights

legislation, as summarised in Tables 8-10.

Regarding Kosovo, since 1999 all international agreements on behalf of Kosovo

were signed by UNMIK. In certain cases Kosovo simply adopted the provisions of an

international instrument into the local legislation. All these conventions are applicable

and valid under Art. 145 of the Kosovo Constitution.

<<Tables 8-10 about here>>

The general trend in legal provisions for human and minority rights that is obvious from

these data is that constitutional human rights provisions are universally present and that

all countries are states parties to the key international conventions. There seems to be

significantly greater reluctance to include minority rights provisions in constitutions or

38

even provide for separate minority rights legislation either at the state or local level.

Moreover, the mere fact of ratifying international instruments and/or adopting

domestic human and/or minority rights legislation says relatively little about the extent

to which any of this translates into measurable improvements in reality. Thus, another

common denominator across the three countries examined here is that of, at times

serious, deficiencies in implementation of existing legislation. Discrimination, and in

some cases even violence, against members of minorities are still widespread, state

funding of minority protection measures remains inadequate, and states often appear

satisfied with token gestures in the form of adopting certain laws, not least to satisfy

international demands, but show little if any enthusiasm for their proper

implementation. Although these problems and challenges apply to most minorities in

the region, the situation of the Roma minorities is particularly problematic.21

The emphasis on individual human rights over (group-specific) minority rights is

consistent with recommendations of power dividers and centripetalists. Liberal

consociationalists, too, appreciate the value of individual human rights provisions, but

their recommendations do not caution against the parallel use of minority rights

provisions, which is consistent with their generally more pronounced emphasis on the

recognition and protection of (self-determined) group identities.

21 For a detailed analysis of existing legislation and implementation see Wolff et al. (2008).

39

Recognition and Protection of Identities: Territory and Population as Boundaries of

Authority

Authority as the legitimate exercise of political power has two boundaries – it is

normally limited to a specific territory and/or a defined group of people. The degree to

which both of these categories shape the boundaries of authority of specific institutions

of government contributes to an assessment of the degree to which group identities are

institutionally recognised and protected.

A central government has the authority to exercise its power within the

territorial confines of the state it is governing and over the residents of this territory

(with the exception of foreign diplomats for example). Some elements of a central

government’s authority may also extend beyond the territorial boundaries of its state,

but then they will normally be limited to that particular state’s citizens, for example in

the field of tax collection. In terms of self-governance regimes, the extent of these two

limitations placed on the exercise of authority is similar. Territorial self-governance

regimes are spatially confined. The powers devolved to them only apply within the

territorial boundaries of the region and, by extension, only to (permanent) residents of

the region. An analogue to authority extending beyond territorial boundaries are

instances of personal autonomy in which the autonomous body has authority over all

individuals belonging to it no matter where they live in the territory of the state or

region concerned.22

22 In reverse, this means that all members of the ethnic group concerned can enjoy the rights accorded to them in the autonomy arrangement anywhere in the territory of the relevant state. This form of autonomy is particularly useful in instances where groups are more dispersed. It is also used to complement territorial forms of autonomy in

40

In Kosovo, the focus is placed on personal autonomy. Based on Art.1 of the

Constitution, Kosovo is defined as a “state of its citizens”. However, Art. 3 states that

Kosovo “is a multi-ethnic society consisting of Albanian and other Communities.” As in

the case of Macedonia, it is a hot political issue whether the majority is also a member

of a community, equally as the others, or is a nation. Nevertheless, communities are

entitled to special rights according to the third chapter of the constitution. A light

asymmetry applies in Kosovo in terms of the local government level. New municipalities

are to be established (mainly in Serb populated areas) which will have more powers

than the other municipalities in areas such as education, health care and culture and the

ability to cooperate on the matters amongst themselves in Kosovo and across borders.

All these municipalities, however, at least in theory, are directly monitored by, and

accountable to, the central government. Therefore, a regional level cannot be

established. However, when analysing the institutional and normative structure of the

new Kosovo state, we have to keep in mind the huge gap between theory and practice

in Kosovo, due to the ongoing boycott of all Kosovo state institutions by Serbs from

northern Kosovo and the establishment and functioning of parallel structures,

supported financially and politically by Belgrade, that govern and regulate almost every

aspect of Serbian parallel society.

In Macedonia, devolution of significant powers to the level of local government

was accompanied by boundary revisions so as to create local self-government units with

over 20% non-majority population, and thus use the special provisions of the post-OFA

specific policy areas (culture, religious affairs, education, etc.) when autonomous territories are ethnically heterogeneous. For a recent theoretical and empirical elaboration of this, see Cordell and Smith (2008).

41

Constitution. In BiH, the recognition and protection of group identities went beyond the

application of mere territorial principles. This is evident in the landmark judgment by

the Constitutional Court of Bosnia and Herzegovina that all constituent nations defined

in the Constitution—Bosniaks, Croats and Serbs (along with Others)—must be

recognized as constituent nations and enjoy equal status as groups, not only at the level

of the state, but also at the level of both entities.

This issue of privileging group identities is one in which liberal consociationalist

views dominate in conflict resolution practice—territorial boundaries of self-governing

entities are self-defined by their populations. The resultant ‘ethnic’ entities run counter

to recommendations made by either centripetalists or power-dividers both of whom

generally prefer administrative, heterogeneous entities to ethnically self-defined ones

with particular group majorities or pluralities.

V. CONCLUSION

If we were to revisit the theoretical debate on different models of conflict resolution in

the light of our three case studies, it would be difficult to assign them clearly to any one

of the three main schools of thought. On the surface, Bosnia and Herzegovina could be

defined as a corporate consociational (con-) federation, Macedonia as a unitary liberal

consociational regime, and Kosovo as a mixed liberal/corporate consociation within a

unitary state. Yet, below this surface, a more complex reality lurks, one that has

significant theoretical implications (see Table 11). What our three case studies

demonstrate in principle, albeit on an empirically yet too narrow basis for broader

42

generalisation, is that the practice of contemporary international conflict regulation

efforts and outcomes does not overlap neatly with any of the three main schools of

thought. Breaking the analysis further down into the three dimensions of institutional

design we can offer a more nuanced assessment on the degree to which specific

prescriptions by advocates of centripetalist, consociational and power-dividing

approaches are reflected in the institutional designs adopted for the regulation of ethnic

conflicts in Bosnia and Herzegovina, Macedonia, and Kosovo.

Thus, we find, first, in terms of state construction that territorial self-governance

is a core feature of institutional design in all three cases. While there is significant

variation in terms of the number and size of self-governing entities, they are

predominantly demarcated on the basis of ethnic homogeneity (municipalities in

Macedonia and Kosovo, Republika Srpska and majority of Federation cantons in Bosnia

and Herzegovina). This is clearly in line with consociational recommendations in the case

of Republika Srpska, while centripetalists would find homogeneity of cantons in the

Federation of Bosnia and Herzegovina, and of municipalities in Kosovo and Macedonia

reflecting their recommendation that relatively compact groups should be divided

across different entities.

Second, as far as the institutions of government are concerned, the findings are

more mixed. In Macedonia and Kosovo, the system is parliamentary, in Bosnia and

Herzegovina it is semi-presidential but with few powers retained at the State level.

Executive power sharing is mandatory in Kosovo and Bosnia and Herzegovina, and has

so far always emerged voluntarily in Macedonia. Legislative power sharing is guaranteed

43

qua qualified voting procedures on issues of ‘vital interest’ in all three countries. The

electoral system for each country’s legislature is a PR-List system, but with some

important modifications. Kosovo uses open lists and has reserved seats for minority

communities for the duration of its first two parliaments under the new constitution.

The limited term under which this corporate consociational feature of reserved seats is

operational is in line with recommendations made by power dividers: to build a ‘sunset

clause’ into consociational arrangements. In Bosnia and Herzegovina, for the time being

at least, separate electoral roles exist for the two Entities, entrenching further the

corporate nature of the consociational arrangements there. In all three cases, these

various dimensions of institutional design are legally entrenched and there is a clear

separation of powers with an independent judicial system. In Bosnia and Herzegovina

regulations are in place that provide for group representation and international judges

in the court system. Similarly, in Kosovo international judges are to be appointed to the

constitutional court for the duration of the mandate of the ICR.

<<Table 11 about here>>

Third, in all three cases, individual and group rights are guaranteed in the constitution

and distinct group identities are recognised across the entire political system of each

state and accorded specific privileges—strongly in line with consociational

recommendations.

Where does this leave us in terms of conflict regulation theory? A first conclusion

that we can draw from the exploration of our three Western Balkan cases is that they all

represent a complex package of power sharing and territorial self-governance. The

44

negotiation of such deals, within a context of international mediation, indicates an

assumption that such complex regimes can contribute to sub-state, state, regional and

international security and stability. In divided societies in which corresponding group

identities have formed and become salient, the degree of self-governance enjoyed by

the different segments of society is often seen as more or less directly proportional to

the level of acceptance of an overall institutional framework within which these

different segments come together and simultaneously need to resolve a range of policy

issues jointly. Self-governance regimes are thus also meant to provide institutional

solutions that allow the different segments of diverse societies to realise their

aspirations for self-determination while simultaneously preserving the overall social and

territorial integrity of existing states, which can only be achieved in a sustainable way if

self-governing communities, and their elites, can meaningfully participate in the political

process at the centre, something which is achieved through power sharing. Thus, the

combination of power sharing and self-governance above all offers institutional

arrangements for conflict parties to settle their disputes by peaceful means.23

Bosnia and Herzegovina, Kosovo and Macedonia, thus, represent three examples

of the notion of complex power sharing,24 which, in the way it is understood here, refers

23 Gurr (1993), even though he did not conceptualize the notion of complex power sharing as such, offered the initial empirical evidence that “some combination of *…+ autonomy and power sharing” offers reasonable prospects to accommodate minority demands (1993:292), a conclusion that is also supported in later work by Weller (2008:xvii), Wolff (1997, 2004; 2008a) and Wolff and Weller (2005). For two empirically broader studies of the complex power sharing phenomenon, see Wolff (2008, 2009a, b), and for a range of single and comparative case studies, Weller and Metzger (2008). 24 We borrow the term ‘complex power-sharing’ from a research project funded by “Project Funder” (“Project Title”). In this project, complex power-sharing regimes are distinguished “in that they no longer depend solely on consociational theory, or solely upon integrative theory”, involve international actors that “are often key in designing, or bringing experience to bear upon, the structure of the eventual agreement, or its implementation” and “consider a far broader range of issues … and … address structural issues as diverse as economic management, civil-military

45

to a practice of conflict settlement that has a form of self-governance regime at its

heart, but whose overall institutional design includes a range of further mechanisms for

the accommodation of ethnic diversity in divided societies, including those

recommended by advocates of liberal consociationalism, centripetalism, and power

dividing. Complex power sharing is thus the result of the implementation of a self-

governance regime whose success as a conflict regulation arrangement requires a

relatively complex institutional structure that cannot be reduced to autonomy/(ethno-

)federalism, (traditional) models of power sharing or power dividing. As we have

demonstrated empirically above, this principle of institutional design has been applied

in all of our Western Balkan case studies, albeit in different concrete manifestations.

Among the three theories of ethnic conflict regulation, (liberal) consociational

features form the core of institutional design in each of our cases, but to different

degrees and complemented by elements of institutional design more in line with

prescriptions emerging from centripetalist and power-dividing theories. Within a liberal

consociational framework, there is room (and a recognized need) for a range of power

dividing strategies, including a strong role for judicial entrenchment and enforcement

mechanisms, and universally applicable and enforceable human rights legislation.

Liberal consociationalism is also open to a vertical division of power on the basis of non-

ascriptive, i.e., non-ethnic criteria but, in contrast to power dividing and centripetalism,

does not rule it out either should self-determined entities on that basis emerge and

relations and human and minority rights, and … do so at many different levels of government”, thus recognising “that at different levels of government, different strategies may be more, or less, applicable, and consequently more, or less, successful, in engendering peace and stability” (Kettley, Sullivan, Fyfe 2001: 4-5). O’Leary (2005a: 34-5) uses the term ‘complex consociation’ in a similar manner.

46

desire territorial self-governance. Similarly, the liberal consociational turn away from

strict group veto powers towards qualified majority voting procedures indicates new

avenues to establish the ‘arenas of bargaining’ recommended by centripetalists to

incentivise political actors from different groups to find sustainable compromises.

Yet, liberal consociationalism is not synonymous with complex power sharing,

even though it offers a promising point of departure for a new research agenda on

conflict regulation theory. Thus, finally, and perhaps most importantly for further

conceptual and theoretical development of the notion of ‘complex power sharing’, the

evidence presented here does not amount to a full-fledged and validated theory of

conflict regulation. For complex power sharing to develop into a theory of its own,

further research is necessary. While we have demonstrated that it describes a particular

phenomenon of conflict regulation practice in adequate detail, more work needs to be

done to increase its predictive capabilities (i.e., when are complex power sharing

regimes likely to emerge) and its explanatory value (i.e., when and why does it succeed).

Only then will it be possible to make sure that complex power sharing does not emerge

accidentally as a patchwork of different conflict regulation mechanisms cobbled

together to accommodate a wide range of diverse (and most likely, incompatible)

interests, but to provide a framework within which stable, lasting and ultimately

successful institutions for ethnic conflict regulation can be designed.

47

REFERENCES Bastian, S. and Luckham, R. (eds) (2003) Can Democracy Be Designed? London: Zed

Books. Benedikter, T. (2007) The World’s Working Regional Autonomies. London: Anthem Press. Chapman, T. and Roeder, P.G. (2007) ‘Partition as a Solution to Wars of Nationalism: The

Importance of Institutions’, American Political Science Review, 101(4): 677-91. Choudhry, S. (ed.) (2008) Constitutional Design for Divided Societies: Integration or

Accommodation? Oxford: Oxford University Press. Co-author (2010) Consociationalism, Power Sharing, and Politics at the Center, in Robert

Denemark et. al. (eds) The International Studies Compendium Project. Oxford: Wiley-Blackwell.

Cordell, K. and Smith, D. (eds) (2008) Cultural Autonomy in Contemporary Europe. London: Routledge.

Cordell, K. and Wolff, S. (eds) (2010) The Routledge Handbook of Ethnic Conflict. London: Routledge.

Darby, J. and McGinty, R. (eds) (2003) Contemporary Peacemaking. Basingstoke: Palgrave.

Fearon, J. (1995) Rationalist Explanations for War. International Organization 49 (3), 379–414.

Fearon, J. (1998) Commitment Problems and the Spread of Ethnic Conflict. In D.A. Lake and D. Rothchild (eds) The International Spread of Ethnic Conflict: Fear, Diffusion, and Escalation. Princeton: Princeton University Press, pp. 107–26.

Ghai, Y. (ed.) (2000) Autonomy and Ethnicity. Cambridge: Cambridge University Press. Guelke, A. and Tournon, J. (2010) (eds) The Study of Politics and Ethnicity: Recent

Analytical Developments. Opladen: Barbara Budrich Publishers. Gurr, T. R. (1993) Minorities at Risk: A Global View of Ethnopolitical Conflicts.

Washington, D.C.: United States Institute of Peace Press. Hartzell, C. and Hoddie, M. (2008) Crafting Peace: Power-sharing and the Negotiated

Settlement of Civil Wars. University Park: Pennsylvania State University Press. Hechter, M. (2001) Containing Nationalism. Oxford: Oxford University Press. Henrard, K. (2000) Devising an Adequate System of Minority Protection. Leiden/Boston:

Martinus Nijhoff. Horowitz, D. L. (1985 [2000]) Ethnic Groups in Conflict. Berkeley, CA: University of

California Press. Horowitz, D. L. (1990) ‘Ethnic Conflict Management for Policymakers’ in Conflict and

Peacemaking in Multiethnic Societies, ed. by J. V. Montville. Lexington, MA: Lexington Books.

Horowitz, D. L. (1991) A democratic South Africa? Constitutional Engineering in a Divided Society. Berkeley: University of California Press.

Horowitz, D. L. (2003) Electoral Systems and Their Goals: A Primer for Decision-Makers, Journal of Democracy, vol. 14, no. 4 (October 2003), 115-27.

Horowitz, D. L. (2004) ‘The Alternative Vote and Interethnic Moderation: A reply to Fraenkel and Grofman’, Public Choice, vol. 121, no.s 3-4 (December 2004), 507-17.

48

Horowitz, D. L. (2006) ‘Strategy Takes a Holiday: Fraenkel and Grofman on the Alternative Vote’, Comparative Political Studies, vol. 39 (June 2006), 652-62.

Horowitz, D.L. (2002) Constitutional Design: Proposals versus Processes. In A. Reynolds (ed.) The Architecture of Democracy. Oxford: Oxford University Press, pp. 15–36.

Horowitz, D.L. (2007) The Many Uses of Federalism. Drake Law Review 55 (4), 953–66. Kettley, C., Sullivan, J. and Fyfe, J. (2001) ‘Self-Determination Disputes and Complex

Power Sharing Arrangements: A Background Paper for Debate’. Cambridge: Centre of International Studies, available online at http://www.intstudies.cam.ac.uk/centre/cps/download/background1.pdf.

Lake, D.A., and Rothchild, D. (1998) Spreading Fear: The Genesis of Transnational Ethnic Conflict. In D.A. Lake and D. Rothchild (eds) The International Spread of Ethnic Conflict: Fear, Diffusion, and Escalation. Princeton: Princeton University Press, pp. 3–32.

Lantschner, E., Marko, J., Petricusic, A. (eds) (2008) European Integration and its Effects on Minority Protection in South Eastern Europe. Nomos: Baden-Baden.

Lapidoth, R. (1996) Autonomy: Flexible Solutions to Ethnic Conflicts. Washington, D.C.: United States Institute of Peace Press.

Lijphart, A. (1977) Democracy in Plural Societies. New Haven and London: Yale University Press.

Lijphart, A. (1995) ‘Self-determination versus Pre-Determination of Ethnic Minorities in Power sharing Systems’, in The Rights of Minority Cultures, ed. by Will Kymlicka. Oxford: Oxford University Press.

Lijphart, A. (2002) ‘The Wave of Power sharing Democracy’ in The Architecture of Democracy: Constitutional Design, Conflict Management and Democracy, ed. by A. Reynolds. Oxford: Oxford University Press.

Lijphart, A. (2004) ‘Constitutional Design for Divided Societies’, Journal of Democracy, vol. 15, no. 2 (April 2004), 96-109.

McGarry, J. (2006) ‘Iraq: Liberal Consociation and Conflict Management’ (Draft working paper, ms. in author’s possession).

McGarry, J. (2007) ‘Asymmetrical Federal Systems’, Ethnopolitics vol. 6, no. 1. McGarry, J. and O’Leary, B. (2004a) The Northern Ireland Conflict: Consociational

Engagements. Oxford: Oxford University Press. McGarry, J. and O’Leary, B. (2004b) ‘Introduction: Consociational Theory and Northern

Ireland’, in The Northern Ireland Conflict: Consociational Engagements, ed. by J. McGarry and B. O’Leary. Oxford: Oxford University Press.

McGarry, J., and O’Leary, B. (2004b) Introduction: Consociational Theory and Northern Ireland. In J. McGarry and B. O’Leary, The Northern Ireland Conflict: Consociational Engagements. Oxford: Oxford University Press, pp. 1–60.

McGarry, J., O’Leary, B., and Simeon, R. (2008) Integration or Accommodation? The Enduring Debate in Conflict Regulation. In S. Choudhry (ed.) Constitutional Design for Divided Societies: Integration or Accommodation? Oxford: Oxford University Press, pp. 41-88.

Noel, S. (ed.) (2005) From Power Sharing to Democracy: Post-conflict Institutions in Ethnically Divided Societies. Montreal: McGill-Queen’s University Press.

49

Norris, P. (2008) Driving Democracy. Cambridge: Cambridge University Press. O’Flynn, I. and Russell, D. (2005) (eds.) Power Sharing: New Challenges for Divided

Societies. London: Pluto Press. O’Flynn, I. and Russell, D. (eds) (2005) Power Sharing: New Challenges for Divided

Societies. Ann Arbour: University of Michigan Press. O’Leary, B. (2010) Debating Partition: Evaluating the Standard Justifications, in K. Cordell

and S. Wolff, eds, The Routledge Handbook of Ethnic Conflict. London: Routledge. O’Leary, B. (2005a) ‘Debating Consociational Politics: Normative and Explanatory

Arguments’, in From Powersharing to Democracy, ed. by Sid Noel. Montreal and Kingston: McGill/Queen’s University Press.

O’Leary, B. (2005b) ‘Powersharing, Pluralist Federation, and Federacy’, in The Future of Kurdistan in Iraq, ed. by B. O’Leary, J. McGarry and K. Salih. Philadelphia: University of Pennsylvania Press.

O’Leary, B., Grofman, B. and Elklit, J. (2005) ‘Divisor Methods for Sequential Portfolio Allocation in Multi-Party Executive Bodies: Evidence from Northern Ireland and Denmark’, American Journal of Political Science, vol. 49, no. 1 (January 2005), 198-211.

O’Leary, B., McGarry, J. and Salih, K. (eds) (2005) The Future of Kurdistan in Iraq. Philadelphia: University of Pennsylvania Press.

Reilly, B. (2001) Democracy in Divided Societies: Electoral Engineering for Conflict Management. Cambridge: Cambridge University Press.

Reynolds, A. (ed.) (2002) The Architecture of Democracy: Constitutional Design, Conflict Management and Democracy. Oxford: Oxford University Press

Roeder, P. G. (2005) ‘Power Dividing as an Alternative to Power Sharing’, in Sustainable Peace: Power and Democracy after Civil Wars, ed. by P. G. Roeder and D. Rothchild. Ithaka, NY: Cornell University Press.

Roeder, P. G. and Rothchild, D. (eds) (2005) Sustainable Peace: Power and Democracy after Civil Wars. Ithaka, NY: Cornell University Press.

Rothchild, D. and Roeder, P. G. (2005) ‘Power Sharing as an Impediment to Peace and Democracy’, in Sustainable Peace: Power and Democracy after Civil Wars, ed. by P. G. Roeder and D. Rothchild. Ithaka, NY: Cornell University Press.

Sambanis, N. (2000) "Partition as a Solution to Ethnic War: An Empirical Critique of the Theoretical Literature." World Politics 52 (July):437-83.

Sambanis, N., and Schullhofer-Wohl, J. (2009) "Is Partition a Solution to Civil War?" International Security 34 (2):82-118.

Schneckener, U. and Wolff, S. (eds) (2004) Managing and Settling Ethnic Conflicts. London: Hurst.

Sisk, T. D. (1996) Power Sharing and International Mediation in Ethnic Conflict. Washington, D.C.: United States Institute of Peace Press.

Taylor, R. (ed.) (2009) Consociational Theory: McGarry & O’Leary and the Northern Ireland Conflict. London: Routledge.

Walter, B. (2001) Committing to Peace: The Successful Settlement of Civil Wars. Princeton: Princeton University Press.

50

Weller, M. (2005) ‘Self-governance in Interim Settlements: The Case of Sudan’, in Autonomy, Self-governance and Conflict Resolution: Innovative Approaches to Institutional Design in Divided Societies, ed. by Marc Weller and Stefan Wolff. London: Routledge.

Weller, M. and Wolff, S. (2005) ‘Recent Trends in Autonomy and State Construction’, in Autonomy, Self-governance and Conflict Resolution: Innovative Approaches to Institutional Design in Divided Societies, ed. by Marc Weller and Stefan Wolff. London: Routledge.

Weller, M., and Metzger, B. (eds) (2008) Settling Self-determination Disputes: Complex Power Sharing in Theory and Practice. Leiden: Martinus Nijhoff.

Weller, M., and Wolff, S. (eds) (2005) Autonomy, Self-governance and Conflict Resolution: Innovative Approaches to Institutional Design in Divided Societies. London: Routledge.

Wilford, R. (ed.) (2001) Aspects of the Belfast Agreement. Oxford: Oxford University Press.

Wimmer, A. (2003) ‘Democracy and Ethno-religious Conflict in Iraq’, Survival, vol. 45, no. 4 (December 2003), 111-134.

Woelk, J., Palermo, F. and Marko, J. (eds) (2008) Tolerance through Law. Leiden/Boston: Martinus Nijhoff.

Wolff, S. (1997) ‘Territorial and Non-Territorial Autonomy as Institutional Arrangements for the Settlement of Ethnic Conflicts in Mixed Areas’, in Contemporary Political Studies 1998, vol. 1, ed. by Andrew Dobson and Jeffrey Stanyer. Nottingham: Political Studies Association of the United Kingdom of Great Britain and Northern Ireland.

Wolff, S. (2003) Disputed Territories: The Transnational Dynamics of Ethnic Conflict Settlement. New York and Oxford: Berghahn.

Wolff, S. (2004) ‘The Institutional Structure of Regional Consociations in Brussels, Northern Ireland, and South Tyrol’, Nationalism and Ethnic Politics, vol. 10, no. 3, 387-414.

Wolff, S. (2005) ‘Electoral Systems Design and Power sharing Regimes’, in Powersharing: New Challenges for Divided Societies, ed. by I. O’Flynn and D. Russell. Ann Arbor: University of Michigan Press.

Wolff, S. (2008a) Complex Power Sharing as Conflict Resolution: South Tyrol in Comparative Perspective. In J. Woelk, F. Palermo, and J. Marko (eds.) Tolerance through Law: Self-governance and Group Rights in South Tyrol. Leiden: Martinus Nijhoff, pp. 329–70.

Wolff, S. (2008b) Power Sharing and the Vertical Layering of Authority: A Review of Current Practices. In M. Weller and B. Metzger (eds.) Settling Self-determination Disputes: Complex Power Sharing in Theory and Practice. Leiden: Martinus Nijhoff, pp. 407–50.

Wolff, S. (2009a) Complex Power Sharing and the Centrality of Territorial Self-governance in Contemporary Conflict Settlements. Ethnopolitics 9 (1), 27-45.

Wolff, S. (2009b) Peace by Design? Towards “Complex Power Sharing.” In R. Taylor (ed.) Consociational Theory: McGarry & O’Leary and the Northern Ireland Conflict. London: Routledge, 110-121.

51

Wolff, S. (2010—forthcoming) The Relationships between States and Non-state Peoples: A Comparative Context for the Kurds in Iraq. In R. Lowe and G. Stansfield (eds.) The Kurdish Policy Imperative. London: Chatham House.

Wolff, S. and Weller, M. (2005) ‘Self-determination and Autonomy: A Conceptual Introduction’, in Autonomy, Self-governance and Conflict Resolution: Innovative Approaches to Institutional Design in Divided Societies, ed. by Marc Weller and Stefan Wolff. London: Routledge.

Wolff, S., van Houten, P., Anghelea, A.-M., Djuric, I. (2008) Minority Rights in the Western Balkans. Report for the European Parliament. Brussels and Strasbourg: European Parliament.

Zartman, I. William (1989) Ripe for Resolution: Conflict and Intervention in Africa. Oxford: Oxford University Press.

52

Table 1: Main Institutional Arrangements Recommended by Different Theories of Conflict Regulation

Centripetalism Liberal Consociational Power sharing

Power dividing

Principal recommendation Interethnic cooperation and moderation induced by electoral system design encouraging vote pooling

Interethnic cooperation at elite level induced by institutional structure requiring jointness of executive decision making

Cooperation between different, changing coalitions of interest induced by separation of powers

Stat

e

Co

nst

ruct

ion

Heterogeneity vs. homogeneity of federal units (if any)

Preference for heterogeneous units

25

Preference for units based on self-determining communities

Preference for heterogeneous units

Number of units relative to number of groups

Preference for more units than groups

Preference for units equal to numbers of groups

No explicit connection between number of groups and units

The

Inst

itu

tio

ns

of

Go

vern

me

nt Government system Presidential

26 Parliamentary or

Collective/Rotating Presidential system

27

Presidential

Executive power sharing Yes: voluntary Yes: guaranteed No, except transition phase after civil wars

Legislative power sharing Yes: voluntary Yes: guaranteed No, except transition phase after civil wars

Electoral system (for parliament)

Plurality preferential PR-List or PR preferential Plurality

Judicial branch Independent Independent and representative

28

Independent

Legal entrenchment Yes Yes Yes

Rig

hts

an

d

Ide

nti

tie

s

Individual vs. group rights

Emphasis on individual rights

Emphasis on combination of individual and group rights

Emphasis on individual rights

Recognition of distinct identities

Yes, but primarily as private matter

Yes, but as private and public matter

Yes, but primarily as private matter

Table 2: Levels of government

Two levels Multiple levels

Macedonia Kosovo

BiH

25 Horowitz is more flexible here than other centripetalists and accepts that sometimes homogeneous units are as useful as heterogeneous units. Personal communication from Donald Horowitz. Cf. also above. 26 Horowitz does not insist that presidential systems are always best and argues that for his recommendations to work it is not essential that governments be presidential. Personal communication from Donald Horowitz. 27 Lijphart is a strong advocate of parliamentary systems, while McGarry and O’Leary are open to other arrangements, insisting that what is crucial for consociation is not whether powers are fused or divided, but whether the different communities are represented in core institutions of the state. This view is supported empirically by the cases of Bosnia and Herzegovina and Switzerland which are both presidential and consociational (personal communication from John McGarry). 28 Liberal consociationalists recognise that it is more difficult to make the judiciary representative than an elected body, but nonetheless note the importance of its representativeness. Personal communication from John McGarry.

53

Table 3: Structural and Functional Symmetry and Asymmetry of Institutions

Structures Functions

Symmetric Asymmetric Symmetric Asymmetric

BiH X X

Macedonia X X

Kosovo x X

Table 4: Distribution of powers between different levels of government

Open-ended list at centre Specific list at centre Mixed list

Macedonia

BiH: State vis-à-vis Entities/Brčko BiH: Federation vis-à-vis Cantons

Kosovo

BiH: Federation—Cantons

Table 5.1: Parliamentary vs. Presidential Systems

Parliamentary system of government Semi-presidential system of government

Macedonia Kosovo

District of Brčko Federation of BiH Republika Srpska

BiH

Table 5.2: Unicameral vs. Bicameral Systems

Unicameral Legislature Bicameral Legislature

Macedonia Kosovo

District of Brčko Republika Srpska

BiH Federation of BiH

Table 6.1: Horizontal Executive Power Sharing at Central and Sub-state Levels of Authority

Power sharing at the centre Power sharing at sub-state level

Macedonia BiH: State

Kosovo

BiH: Federation of BiH/Cantons BiH: District of Brčko

Kosovo Macedonia

Table 6.2: Representation vs. Participation in Power Sharing Arrangements

Mandatory representation in executive and/or legislature

Qualified majority voting procedures

BiH: State BiH: Federation of BiH/Cantons

BiH: District of Brčko BiH: Repubika Srpska

Kosovo: State/Municipalities

BiH: State BiH: Federation of BiH/Cantons

BiH: District of Brčko Kosovo: State

Macedonia

Table 7: Guarantees of Self-governance Institutions

International Guarantees Domestic Guarantees

‘Hard’ ‘Soft’ Constitutional Guarantees Guarantees in Specific Laws

BiH: all levels Kosovo

Macedonia

BiH: all levels Macedonia

Kosovo

Macedonia Kosovo

54

Table 8: International Human and Minority Rights Instruments

UN

Me

mb

ers

hip

Co

nve

nti

on

on

th

e P

reve

nti

on

an

d

Pu

nis

hm

en

t o

f th

e C

rim

e o

f G

en

oci

de

Inte

rnat

ion

al C

on

ven

tio

n o

n t

he

Elim

inat

ion

of

All

Form

s o

f R

acia

l

Dis

crim

inat

ion

Inte

rnat

ion

al C

ove

nan

t o

n E

con

om

ic,

Soci

al, a

nd

Cu

ltu

ral R

igh

ts

Inte

rnat

ion

al C

ove

nan

t o

n C

ivil

an

d

Po

litic

al R

igh

ts

Op

tio

nal

Pro

toco

l to

th

e In

tern

ati

on

al

Co

ven

ant

on

Civ

il an

d P

olit

ica

l Rig

hts

UN

ESC

O M

em

be

rsh

ip

UN

ESC

O C

on

ven

tio

n a

gain

st

Dis

crim

inat

ion

in E

du

cati

on

BiH Yes 1992 1993 1993 1993 1995 Yes 1993

Macedonia Yes 1994 1994 1994 1994 1994 Yes 1997

Kosovo No No 2001 No 2001 2001 No No

Table 9: Regional Human and Minority Rights Instruments: Europe

Council of Europe Membership

(European) Convention for the Protection of Human Rights and Fundamental Freedoms

(European) Framework Convention for the Protection of National Minorities

European Charter for Regional or Minority Languages

BiH Yes 2002 2000 (2005)29

Macedonia Yes 1997 1998 (1996)

30

Kosovo No 2001 2001 2001

Table 10: Domestic and Local Human and Minority Rights Provisions

Co

nst

itu

tio

nal

Hu

man

Rig

hts

Pro

visi

on

s

Co

nst

itu

tio

nal

Min

ori

ty R

igh

ts

Pro

visi

on

s

Stat

e-w

ide

Min

ori

ty

Rig

hts

Le

gisl

atio

n

Loca

l Hu

ma

n R

igh

ts

Legi

slat

ion

Loca

l Min

ori

ty R

igh

ts

Legi

slat

ion

BiH Yes Yes Yes Yes Yes

Macedonia Yes Yes Yes N.A. N.A.

Kosovo

Yes Yes Yes Yes Yes

29 Not yet ratified. 30 Not yet ratified.

55

Table 11: Institutional Arrangements in Bosnia and Herzegovina, Macedonia, and Kosovo

Bosnia and Herzegovina Macedonia Kosovo

Principal design Corporate consociational con-federation

Liberal consociational unitary state

Mixed liberal-corporate consociational unitary state

Stat

e C

on

stru

ctio

n

Heterogeneity vs. homogeneity of federal units (if any)

Federation: heterogeneous, but with more homogeneous cantons RS: homogeneous Brčko: heterogeneous, but highly segregated

Unitary state, with highly homogeneous local municipalities

Unitary state, with highly homogeneous local municipalities

Number of units relative to number of groups

One principal unit for Serbs, multiple for Croats and Bosniaks

More units than groups More units than groups

The

Inst

itu

tio

ns

of

Go

vern

me

nt

Government system Semi-presidential with rotating collective presidency

Parliamentary Parliamentary

Executive power sharing

Yes: guaranteed Yes: voluntary Yes: guaranteed

Legislative power sharing

Yes: guaranteed for ‘vital-interest’ legislation

Yes: guaranteed for ‘vital-interest’ legislation

Yes: guaranteed for ‘vital-interest’ legislation

Electoral system (for parliament)

Closed PR-List, separate electoral roles for Entities

Closed PR-List Open PR-List with reserved seats for ‘communities’

Judicial branch Independent and representative (with international judges in constitutional court)

Independent Independent (with international judges in constitutional court)

Legal entrenchment Yes Yes Yes

Rig

hts

an

d

Ide

nti

tie

s

Individual vs. group rights

Individual and group rights guaranteed in constitution

Individual and group rights guaranteed in constitution

Individual and group rights guaranteed in constitution

Recognition of distinct identities

Yes Yes Yes